Article type
Year
Abstract
Objective: To provide experience in applying AMSTAR2 to systematic reviews (SRs) that include non-randomized studies of interventions (NRSI).
Description: SRs of studies of healthcare interventions increasingly include both RCTs and NRSI. AMSTAR, a popular critical appraisal tool, was developed to evaluate SRs of RCTs. We aimed to adapt AMSTAR for appraisal of reviews that include NRSI. We recruited an expert panel who received results of an updated literature review and a survey of user experience of the original AMSTAR. The panel recommended that the revised AMSTAR be modified to have: better alignment with the PICO (population, intervention, comparator, outcomes) framework; clearer justification of the selection of study designs for inclusion in SRs; stronger justification for exclusion of studies from SRs; an assessment of whether reviewers had made a sufficiently detailed assessment of risk of bias (RoB, using guidance from ROBINS-I) for the included studies (both RCTs and NRSI). The resulting instrument has 15 items, four of which address different aspects of bias in the studies included in SRs. We will provide the instrument and reviews to participants, who will work in groups. The workshop will be guided by facilitators involved in the development of AMSTAR2, who have experience in the assessment of RoB with individual NRSI. Workshop participants will provide information that will be very valuable in the continuing development of the instrument.
Description: SRs of studies of healthcare interventions increasingly include both RCTs and NRSI. AMSTAR, a popular critical appraisal tool, was developed to evaluate SRs of RCTs. We aimed to adapt AMSTAR for appraisal of reviews that include NRSI. We recruited an expert panel who received results of an updated literature review and a survey of user experience of the original AMSTAR. The panel recommended that the revised AMSTAR be modified to have: better alignment with the PICO (population, intervention, comparator, outcomes) framework; clearer justification of the selection of study designs for inclusion in SRs; stronger justification for exclusion of studies from SRs; an assessment of whether reviewers had made a sufficiently detailed assessment of risk of bias (RoB, using guidance from ROBINS-I) for the included studies (both RCTs and NRSI). The resulting instrument has 15 items, four of which address different aspects of bias in the studies included in SRs. We will provide the instrument and reviews to participants, who will work in groups. The workshop will be guided by facilitators involved in the development of AMSTAR2, who have experience in the assessment of RoB with individual NRSI. Workshop participants will provide information that will be very valuable in the continuing development of the instrument.